Microsoft Translator Hub: Discussion of BLEU Score
Chris Wendt interviews Dr. Will Lewis to get a deeper insight into automatic scoring of customized machine translation systems. Microsoft Translator Hub & Bing Translator use an automatic scoring mechanism known as BLEU (Bilingual Evaluation Understudy). Dr. Will Lewis explains how this scoring mechanism works, how it compares to human evaluation and what factors impact BLEU score.
- Date:
- Haut-parleurs:
- Chris Wendt and Will Lewis
-
-
Chris Wendt
Principal Group Program Manager
-
Lori Stone
-
Will Lewis
Principal PM Architect
-
-
Regardez suivant
-
-
Microsoft Research India - who we are.
Speakers:- Kalika Bali,
- Sriram Rajamani,
- Venkat Padmanabhan
-
Multilingual Evaluation of Generative AI (MEGA)
Speakers:- Kabir Ahuja,
- Stephanie Nyairo,
- Millicent Ochieng
-